Abstract:

Typical B-R assessment involves separate data summaries for each efficacy and safety outcome. For example in clinical trials, outcome-specific effects are estimated and potentially systematically or unsystematically combined in B-R assessment with the belief that such analyses inform the totality of effects on patients. However, this approach is suboptimal for informing medical decision-making, as it fails to: (i) evaluate the association between component outcomes (i.e. efficacy, toxicity, quality of life); (ii) evaluate the cumulative nature of outcomes; (iii) systematically incorporate the relative importance of the component outcomes; (iv) address competing risks, and since efficacy and safety analyses are conducted on different analysis populations, the population to which these B-R analyses apply, is unclear. In this Webinar, we discuss these issues and describe methods to address them.  The Desirability of Outcome Ranking (DOOR) is an approach to pragmatic benefit:risk evaluation to better inform medical decision-making . When applying these methods in a clinical trial for example, the probability of more desirable overall result from a B-R perspective when using a treatment relative to a control, can be estimated. This probability may be attractive for clinicians and patients when making treatment decision in practice. Partial credit can also be used to incorporate patient/clinician preferences associated with global patient outcome. Visual displays can be used to display treatment contrast as the partial credit preferences vary. The methods are illustrated using examples in infectious diseases and neurology.